►
From YouTube: Numenta Research Meeting - Apr 24, 2019
Description
Broadcasted live on Twitch -- Watch live at https://www.twitch.tv/rhyolight_
A
C
C
B
A
B
A
So
so
here,
like
the
the
most
succinctly,
I,
can
give
of
this
that
the
clearest
one
is
what
this
represents
is,
if
you
take
your
current
plane
and
rotate
it
toward
the
gravity
vector.
This
will
point
in
your
current
direction
that
this
tells
you
what
direction
you'll
be
pointing
if
you
take
your
current
plane
and
rotate
it
to
the
gravity
vector
so,
okay.
A
D
A
C
A
D
C
C
C
C
Me
of
sort
of
like
we
talked
about
being
on
an
object
and
not
knowing
the
location,
not
knowing
the
orientation
now
I'm
at
a
point
in
me
in
the
the
orientation
space
and
I'm,
not
changing
my
point
in
orientation
space,
but
I'm
rotating
around,
so
they
seem
qualitatively
different.
You
know
I'm,
saying
like
if
I
was
thinking
about
what
I
see
in
a
room,
the
object
I'm
seeing
is
going
to
be
determined
by
the
2d
space,
but
the
actual
orientation
of
that
presentation
of
my
would
would
has
to
be
determined
by
that
guy.
C
A
B
C
But
that
doesn't
mean
that
that
doesn't
mean
that
that's
not
changing
right
I
mean
if
you're
on
on
that
sphere,
the
plane
with
a
rat
on
the
sphere
there
right
every
it
so
hard
to
translate
that
to
real
world,
but
I
mean
the
fact
that
you're
changing
in
the
2d
space
does
not
mean
that
one
day
space
won't
be
changed.
Well,.
A
C
Alright,
so
that
interpretation
was
wrong.
My
my
interpretation
wrong.
That's
not
a
good
way
to
think
about
it.
It's
kind
of
machi
I
think
about
this
problem.
I
find
a
little
bit
difficult
to
just.
Maybe
that's
what
you're
trying
to
do
it
internalize
what
this
means.
You
know
I!
Look
at
that
and
imagine
oh
that's
my
guess
what
I'm
pointing
around
you
know
my
position.
C
B
C
A
C
D
C
B
E
E
C
C
Did
then
I'll
be
very
confident
that
it's
correct
at
this
point,
I
was
just
trying
out
an
idea
and
without
even
any
implementations
for
the
biology,
so
I
I'm
not
disagreeing
with
you,
but
it
was
like
when
we
did
the
sequence
memory
and
all
that
stuff
I
felt
like
I
was
gonna
work
and
other
people
say.
Well.
Maybe
we
don't
really.
C
C
B
C
D
E
E
B
E
B
E
B
E
D
D
C
E
C
C
E
C
D
E
E
E
B
C
C
E
B
E
B
E
C
It
more
tractable:
what
does
this
do
that
this
is?
Is
it
have
the
same
sort
of
MCAS
so.
E
E
D
B
E
C
E
B
E
E
C
B
C
C
E
E
E
E
D
E
B
E
C
E
E
E
E
E
D
E
So
I'm
implement
I've,
not
implemented
the
other
types
in
here
and
I'm
playing
around
with
it.
It's
just
very
complicated
the
training,
the
other
part
of
it,
is
I'm
stepping
back
just
to
look
at
how
sparsity
is
working
in
just
one
of
these
layers,
just
on
its
own,
so
I
can
understand
it
better,
rather
than
trying
to
just
do
a
full
black
box,
end-to-end
kind
of
training,
so.
C
C
E
B
C
Saw
like
it
seems
a
bit
like
the
idea
of
adding
Farsi
becomes
a
complicated
thing,
because
these
these
networks
has
its
own
sort
of
architecture
and
tweaks
that
they
do
and
and
depending
on
which
architecture
you
plan
on
the
sparsity.
Adding
questioning
could
be
a
separate
challenge.
Is
that
correct,
I.
E
D
C
C
B
E
C
C
Yeah
one
of
the
blocks
of
the
thing
you
get
a
block
to
work,
it'll
be
good,
I
mean
again,
you
know
what
else
do
you
want
to
people
to
try
this
very
quickly?
You
know.
Ultimately,
if
you
want
right,
people
said:
hey,
I'm,
gonna
drop
this
into
my
network,
see
what
happens
and
it's
the
extended
amount
of
effort
they
have
to
go
through
to
do
that
is
directly
counting
people.
Try
that.